[PS-3.10] Challenging the distinction between presupposition holes and presupposition plugs

Patel-Grosz, P. , Holweger, M. , Kiss, N. & Jaeger, G.

Eberhard Karls University of Tuebingen

In (1), the factive predicate 'be glad' triggers a presupposition (that Bill is in town). In an offline ratings experiment (40 items in 2x2 design), we tested whether native speakers of German are sensitive to the following distinction. It is assumed (Karttunen 1973) that presuppositions project past operators (so-called "holes"); other operators block them (so-called "plugs"). Participants read a German sentence that contained a hole (e.g. 'Perhaps' as in (1)) vs. a plug (e.g. 'Markus told Hanna that?'). Subsequent questions asked about the certainty (7-point scale) of the presupposed proposition in the global context, (2), or in the local context (prefacing (2) with 'Assuming that this conjecture is correct?').
(1) Perhaps Jonas is glad that Bill is in town.
(2) How certain is it that Bill is in town?
The traditional view predicts that the [hole+global] condition is rated higher than the [plug+global] condition. Our experiment challenges this view by showing that participants (n=45) are not sensitive to the hole/plug distinction. We find a main effect of question type (p<0.001), with higher ratings for local questions than for global questions. We do not find a main effect of hole/plug or an interaction. Average ratings: [hole+local]=5.78, [plug+local]=5.72, [hole+global]=5.38, [plug+global]=5.42